This week's signal is control of the rails. The Compass–MLS fight turns from ultimatums into lawsuits: Compass says it will sue Austin's Unlock MLS after California's CRMLS sued first, and HomeServices of America CEO Chris Kelly calls the entire listing-distribution fight a 'manufactured crisis' — arguing it was never about seller choice, but about monetization and controlling where consumers look for listings. Meanwhile the infrastructure bets run in the other direction: eXp stops building another proprietary chatbot and plugs agents' own AI into its production data via MCP; ZipISA sells its seller-prospecting AI to one brokerage per ZIP, turning exclusivity itself into the product; and Planpoint comes out of stealth turning floor plans into interactive, analytics-backed sales tools. Different plays, same question: who owns the layer everyone else builds on?
01
Make it shine
Tools that make every listing look irresistible.
1 story
New3D & SpatialFrom $15/mo for 5 listings
Planpoint exits stealth, turning floor plans into interactive sales tools
On October 8, Montréal-based Planpoint emerged from stealth with a platform that turns floor plans and site maps into interactive tools for selling and renting property. Prospective buyers and renters can explore available units in context — interactive stacking plans and site maps, photos pinned to the room on the floor plan, saved favorites and floor-plan downloads — while property teams publish straight onto their own websites via embeds for WordPress, Wix, Webflow, or plain HTML, update prices and availability without waiting on a programmer, run touchscreen sales presentations on TVs and tablets, and track engagement (unit views, floor-plan downloads, favorites). The company says the platform has recorded 24.7 million unit views, 15.3 million floor-plan downloads, and 1 million favorites; named customers include Brigil, Cogir, and Mach. Commercial Observer, covering the launch, notes the company was founded in 2018, spent four years building its JavaScript-based software, and is self-funded. Pricing starts at $15 per month for five active listings. All usage and customer figures are company-reported, not independently verified.
Why it matters
Floor plans are the listing visual buyers actually use to evaluate a home — and they mostly still sit as static PDFs. The practical wins here are mundane, which is the point: no programmer to update availability, embed-anywhere publishing, and real engagement analytics on which units draw interest. The caveat is audience fit: Planpoint's named customers are developers and property teams, so resale agents should trial the interactive embeds against simpler options like CubiCasa before committing. Watch the pattern: visuals you can measure are replacing visuals you can only admire.
Tools that put every listing in front of the right eyes.
2 stories
UpdateMarketing Stack
Compass says it will sue Unlock MLS — the second front of the listing-rules war
On October 7, one day after the deadline in its September 8 demand letter, Compass confirmed to Real Estate News that it plans to sue Austin's Unlock MLS, which had refused to stop fining agents who publicly market office-exclusive listings. Unlock CEO Emily Girard said the demand 'strikes at the very structure of the MLS,' arguing the rules keep the marketplace complete, accurate, and equitable; the MLS's outside counsel accused Compass of praising Unlock's Flex private-marketing program as a 'model of seller choice' in March, then threatening an antitrust lawsuit over the same rules six months later. The practical detail from Unlock's response: a seller can already exclude a listing from mass MLS dissemination by signed certification and timely payment of a flat $100 fine — modest enforcement of a mandatory NAR rule — which suggests Compass is litigating the principle, not the price. Separately, Compass said it would file its suit against California's CRMLS 'in the next few weeks,' following CRMLS's October 5 preemptive lawsuit asking a federal court to declare its cooperation rules lawful. All party statements and legal characterizations are reported positions; legal outcomes are unknown.
Why it matters
Direct follow-up to Issue #006's CRMLS-sues-first story — the fight is now on both coasts. For agents, the Unlock detail matters most: even a flat $100 fine with signed seller certification wasn't acceptable to Compass, so this fight is about who gets to set the rules, not how much the fines are. Practical move this month: read your own MLS's rule text and fine schedule, because the distribution options you sell around may be rewritten by judges next year.
HomeServices CEO calls the listing fight a 'manufactured crisis'
In an October 9 interview with Inman, conducted ahead of HomeServices of America's acquisition of real-estate strategy firm 1000WATT earlier this month, CEO Chris Kelly called the industry's year-long fight over listing distribution and private inventory a 'manufactured crisis.' His argument: sellers were not, in his telling, 'forming picket lines' demanding less exposure — the debate is really about monetization and 'controlling where consumers look for listings,' with terms like 'seller choice' and 'three-tier marketing' invented to make it more sellable. Kelly added that the industry had largely already figured out how to handle sellers who genuinely need limited marketing before the current debate erupted. On the company side, he framed 2026's moves — the OnePoint brokerage-mortgage-title-insurance integration, the Cotality Broker Listing Exchange (BLX) partnership for centralized listing entry and distribution, in-house mortgage servicing via Prosperity Home Mortgage, and the 1000WATT acquisition — as turning HomeServices from a decentralized holding company into an active national parent. All characterizations of the dispute are Kelly's opinions, not verified claims.
Why it matters
Read this as the counter-brief to the litigation story. Kelly's framing — that 'seller choice' language is packaging for an eyeball-capture business model — is the lens through which to read every portal and brokerage announcement for the next year. The practical question it hands agents is sharp: for your actual sellers, who benefits from limited distribution — the seller, or the brokerage whose site keeps the traffic? Note the parallel bet: HomeServices is quietly building the same thing the fight is about — centralized listing entry and distribution via BLX.
Turn attention into relationships, and relationships into commissions.
2 stories
NewMarketing Stack
eXp stops building a chatbot, plugs your own AI into its data instead
On October 8, at its eXpcon conference in Salt Lake City, eXp Realty unveiled 'MCP First, AI of Choice' — a Model Context Protocol strategy that connects the AI assistant agents already use (Claude, ChatGPT, or whichever they reach for) directly into their eXp business data: production numbers, leads, transactions, training progress, and CRM records, all from inside the same chat window with no new app or login. Launch integrations cover The Hub, My eXp, eXp University, eXp Divisions, and Canva, plus CRMs Lofty, BoldTrail, and Cloze (with Follow Up Boss and Sisu marked as coming soon). The company pitches it as the opposite of the closed-chatbot race: CEO Leo Pareja said 'MCP is the front door, and AGNT OS is the whole house,' describing AGNT OS — the rebuild of the systems agents use to close deals and get paid, built on the Nexus platform — as targeted for early 2027. Notable thread: Lofty, on the launch CRM list, shipped its own MCP server in September (Issue #005) — the open-connection pattern is spreading. All figures, timelines, and capability claims are company statements carrying forward-looking caveats, not independently verified.
Why it matters
The most structurally interesting AI announcement in weeks — and it lands opposite to every other brokerage's chatbot. Instead of locking agents into one proprietary assistant, eXp is betting the scarce asset is the data connection, not the model — a quiet admission that agents will use whatever AI they want. If you're at eXp, the test is what your AI can actually see and do inside those data stores, and which CRM policies gate it. For everyone else: if your brokerage builds a closed assistant while eXp opens the pipes, note where the agent goodwill lands.
ZipISA sells its seller-prospecting AI to one brokerage per ZIP
On October 12, AI seller-prospecting platform ZipISA announced it is selling its system to brokerages — with a catch: only one brokerage brand per ZIP code. Each package covers three ZIPs and nine agent seats, and once a brokerage claims its ZIPs, ZipISA says it will sell no new seats there to other brokerages or individual agents. The product puts an AI agent on each seat running 100+ seller workflows: it finds likely sellers from signals including expired listings, intent data, named website visitors, and home-value requests; writes and sends emails in the agent's name with follow-ups on days 3, 7, and 14; answers replies; and books calls. New contacts sync into Follow Up Boss, Lofty, or kvCORE/BoldTrail. Since its June 9 launch, the company claims its AI found 160,000+ likely sellers across nearly 5,000 ZIPs, contacted 80,000+, sent 155,000+ emails, passed nearly 19,000 contacts into agents' CRMs, and booked about 150 calls, with about 250 agents using it today. The company — built by AI venture studio Standard and a member of the NVIDIA Inception program — frames it as 'we get agents the at-bats; winning the listing is still their job,' and runs a live webinar for brokerage owners on October 13 at 1pm PT. All performance figures are vendor-reported and unverified; this is a company-issued press release, not independent coverage. Note the timing: the webinar lands the day after this issue's drafting, so the date may be stale by publication.
Why it matters
Look past the webinar and see the pricing model of power: ZipISA isn't selling software, it's selling scarcity — one ZIP, one brokerage. That's a franchise-style land grab dressed as AI tooling, and it inverts the usual SaaS play where the tool goes to everyone. Questions to ask before being enchanted: who defines a 'likely seller,' how are contact consent and email deliverability handled, and what happens when a rival offers a better machine next quarter? Read the vendor numbers as marketing, not evidence — 19,000 CRM contacts against 155,000 emails is the ratio to interrogate.
97% of brokerage leaders say their agents use AI — and expansion plans move it past marketing
Delta Media Group's 2026 Real Estate AI & Leadership Survey, reported by Florida Realtors in its October 2026 issue and based on more than 100 brokerage leaders, is the newest read on where brokerages plan to take AI. 97% of brokerage leaders said their agents use AI. Looking ahead, leaders plan to adopt or expand AI for customer management and workflows (55%), administrative and back-office automation (53%), and agent recruiting, training, and performance coaching (52%) — plus business intelligence and forecasting (49%), document summaries and contract review (48%), and market analysis and valuations (46%). Half plan to adopt or expand agentic AI, the first year the survey asked about it. Leaders rated AI's importance to their businesses 7.12 out of 10, up from 5.45 in 2024, and expect 8.25 over the next 12–18 months. Timing note: the survey itself was released September 16 — just before this issue's research window (October 8–12) — and is included as the freshest agent-AI dataset available. Note also that Delta Media Group is a real-estate technology company, and these are survey self-reports, not independently audited figures.
97%
of brokerage leaders say their agents use AI (survey of 100+ brokerage leaders)
7.12/10
leaders' rating of AI's importance to their business, up from 5.45 in 2024 — expected to reach 8.25 within 12–18 months
55%
plan to adopt or expand AI for customer management and workflows; 53% for admin and back-office automation
50%
plan to adopt or expand agentic AI — the first year the survey asked about it
“Nobody is fighting over homes anymore. They're fighting over the rails the homes move on.”
Every story this week is infrastructure disguised as something else. Compass and the MLSs argue about seller choice, but the litigation is about who controls distribution. HomeServices' CEO says the quiet part out loud: it was never really about sellers — it was about eyeballs. eXp's MCP bet concedes the chatbot race and fights over the data pipes instead. ZipISA doesn't sell AI; it sells ZIP-code exclusivity. Planpoint turns a static floor plan into a measurable channel. The pattern is clean: value has moved from the asset to the rails — from the listing to everything that decides who sees it, who follows up, and who gets paid.
This week, list the rails your business runs on: your MLS's rules and fines, your portal distribution, your CRM's AI integrations, your lead sources. For each one, ask the Compass question — who actually controls it, and what happens if they change the terms tomorrow?
Editorial disclosure: NanaStack is independent. Nothing in this issue is sponsored. Vendor and company claims are reported from the cited sources and are not independently tested: Planpoint's usage figures (24.7M unit views, 15.3M floor-plan downloads, 1M favorites) and pricing (from $15/mo for five listings) are company statements covered by Commercial Observer; eXp's MCP integration list, AGNT OS timeline (early 2027), and all capability claims are company statements carrying forward-looking caveats; ZipISA's performance figures (160,000+ likely sellers, 80,000+ contacted, 155,000+ emails, ~19,000 CRM contacts, ~150 calls, ~250 agents) are vendor-reported in a company-issued press release. Delta Media Group's survey figures are self-reports from 100+ brokerage leaders collected by a real-estate technology company, reported via Florida Realtors; the survey was released September 16, 2026 — just before this issue's research window (October 8–12) — and is included as the freshest agent-AI dataset available, as dated in the text. Compass/Unlock/CRMLS statements are reported positions from company spokespeople and counsel letters; legal outcomes are unknown. Chris Kelly's characterizations of the listing-distribution dispute are his opinions, not verified claims. The ZipISA webinar referenced (October 13, 1pm PT) was announced in the source and may be stale by the time this issue publishes; details were correct at drafting.
